How cannabis use affects alcohol treatment outcomes Introduction: While some suggest that cannabis use is related to increase drinking among individuals with alcohol use disorders (AUD), others have found that cannabis may substitute for alcohol and therefore reduce alcohol consumption among these individuals. However, no study has prospectively examined whether cannabis use is associated with increases or decreases in alcohol use in individuals in treatment for AUD. Primary aims: The primary aims of this study are to examine the effects of changes in cannabis use (i.e., increased vs. decreased use) during AUD treatment on post-treatment (immediately after treatment; one year, three years post-treatment) alcohol use and related problems. We will also investigate potential moderators (e.g., baseline dependence severity, social support) and mediators (e.g., alcohol cravings) of the relationship between cannabis use and AUD treatment outcomes. Preliminary analyses will also examine differences between those who use vs. do not use cannabis. Design: The proposed study is a secondary analysis of the two largest AUD treatment studies undertaken in the United States: the COMBINE Study and Project MATCH. Methods: Our sample consists of individuals with AUD who use cannabis but are not polysubstance dependent. Analyses will employ novel, highly sophisticated statistical methods, including Targeted Maximum Likelihood Estimation (TMLE) and Super Learner. TMLE and Super Learner use machine-learning and cross-validation to control for confounding and reduce bias; these features are particularly important because we are not able to randomize the cannabis use exposure and must control for selection factors. Hypotheses: We hypothesize that increased cannabis use during AUD treatment predicts increased drinking and related problems post- treatment compared to decreased cannabis use. We further hypothesize that high baseline alcohol use severity and low baseline social support worsen the negative effects of cannabis use on alcohol outcomes. Finally, we hypothesize that the effects of cannabis on alcohol use are mediated by increases in alcohol cravings. Innovation: This study will be the first to 1) quantify the long-term effects of cannabis use among individuals in treatment for AUD; 2) examine moderators of the effects of cannabis use on alcohol-related outcomes in individuals in treatment for AUD; and 3) assess mechanisms by which cannabis use affects alcohol-related outcomes in these individuals. Results will inform the clinical ambiguity regarding cannabis use and treatment for AUD, which is particularly timely given that the recent decriminalization of cannabis in several US states may increase the frequency or severity of cannabis use in this population.